Image source: Generated by Unbounded AI

In the magical world of "Harry Potter", the Flying Spell can summon any item the wizard needs. As a highly practical and frequently used spell, the "Flying Curse" appears frequently in books. The protagonist Harry Potter once used the "Flying Curse" to summon a firebolt and successfully passed through the guard of the fire dragon in the first event of the Triwizard Tournament; he also used the "Flying Curse" to summon a flying broomstick and rushed to Hogg. Woz joins in the rescue. Back in the real world, a group of e-commerce platforms and AI startups are trying to make the "Feilai Spell" a practical tool to help merchants quickly match supply sources or help users shop conveniently. Recently, Alibaba International launched the world's first AI search engine in the B2B field, and its name is Accio in English. It can do this. After merchants input their own needs, they will give product recommendations through AI. Immediately afterwards, the foreign AI search engine Perplexity also recently launched the "Buy with Pro" shopping function. Users also only need to enter their needs, and the platform will provide AI-organized product information. If we go back further, in August this year, Amazon also launched AI shopping assistant Rufus on the site. Users can ask questions to Rufus, and Rufus will give corresponding answers based on product details, user reviews and other information on the platform. The new combination of AI search + e-commerce has quietly made its debut and is growing rapidly behind the scenes of the industry. According to the latest data from research institution InsightAce Analytic, the market size of AI e-commerce solutions will soar from US$6.01 billion in 2023 to US$19.89 billion in 2031, with a compound annual growth rate of 16.2%. In the past, traditional shelf e-commerce shopping scenarios, consumers could only enter limited keywords and then filter by themselves among the links pushed by the platform. However, as the supply of products increases, the selection process becomes more time-consuming and labor-intensive, and the search model gradually fails to keep up with users' needs for finding products. Nowadays, when AI and e-commerce business are gradually integrated and AI search engines learn to recite the "flying curse", will consumers' favorite products really be delivered to them in one second?

When AI becomes an agent and shopping guide

Accio launched by Alibaba International this time is open to global merchants, targeting B-end customers engaged in global trade. According to the official introduction of Alibaba International: "Due to the complexity of global trade, merchants used to go through 28 steps from searching for products to finalizing the transaction." In this process, in order to find suitable suppliers, merchants first need to go through different suppliers. Price comparisons and negotiations between merchants; coupled with the large regional span of global trade, merchants also need to pay close attention to changes in logistics prices and foreign exchange rates. Therefore, in the past, merchants who wanted to engage in global trade usually hired overseas experts or agencies at high prices. Accio aims to help merchants save this cost and simplify the global trade process that originally required 28 steps to be completed to 3 steps.

From the appearance point of view, Accio is better than the traditional BThe 2B e-commerce platform is more like an AI dialogue application. In the middle of the screen is a dialog box for merchants to input requirements. Using Accio to find suppliers, the first step for merchants is to enter their product requirements in the dialog box; the second step is to find the best choice after obtaining the product encyclopedia information provided by Accio with the help of AI analysis; the third step Then click on the link of the relevant product to complete the purchase. These three steps respectively correspond to Accio's three major modules, namely:

AI-driven search. Accio will use AI technology to disassemble and clarify merchants' purchasing needs, and support merchants to enter text, pictures or product links; AI reconstructed product encyclopedia page, given by Accio The answer is not a simple link, but an AI encyclopedia end-to-end e-commerce platform that includes prices, purchase quantity limits, and product characteristics. Accio has products from Alibaba International Station, various independent stations around the world, third-party websites and even offline factories. Hundreds of millions of suppliers can help merchants complete end-to-end procurement.

Accio started internal testing in September this year and was officially launched and put into use in November. In an interview with the media, Zhang Kuo, President of Alibaba International Station, once said: "The current version of Accio is only a 'semi-finished product', which only includes 10% of our imagination of the B2B business form in the AI ​​era, but it has already shown a certain degree of potential. Capability. "It is understood that in previous internal tests, with the help of Accio, merchants' conversion rate from search to purchase inquiry has been directly improved by 20-30%.

If Accio plays the role of more like an intermediary agent in global trade, then Perplexity, a much-anticipated AI search platform, is trying to turn AI into a shopping guide that can be seen everywhere when we shop in offline malls. . In November this year, Perplexity AI launched the member shopping function Buy with Pro. The specific usage steps are similar to Accio. You only need to enter your requirements, AI will give product selections, and then click "Buy with Pro" to place an order and pay. Perplexity also launched the Snap to Shop function, which also supports image search. Users only need to upload images to get detailed product information and purchase links compiled by AI. Since Perplexity AI is just an AI search engine company and does not have huge product resources like Alibaba International, Perplexity chose to cooperate with Shopify, currently the largest independent website platform, and integrate independent website merchants on the Shopify platform. At the same time, Perplexity has also launched a merchant program, inviting brands and merchants to input their product data into Perplexity’s AI system. In this way, brands and merchants will be able to obtain new sales channels, and Perplexity will also be able to provide Provide users with a richer product supply. The biggest difference between the product services provided by Perplexity and traditional e-commerce platforms such as Amazon is that Buy with Pro currently does notRely on advertising for profit. When traditional e-commerce platforms recommend products, they usually sort them according to the price of merchant advertising, while the product recommendation results given by Perplexity are entirely judged by AI. Sometimes even the founder of Perplexity has a hard time explaining the specific AI judgment logic behind a certain answer. This is just like when you walk into a store and ask the shopping guide "Christmas is coming soon, what gift is better for your parents?" Different shopping guides will recommend different products to you based on their own experience. The way of asking and the details of your needs The degree of personalization may affect the judgment of the AI ​​shopping guide.

From "My Daily" to "My Shopping Mall"

In 1995, writer Negroponte proposed "My Daily" in his book "Digital Survival" to describe personalized news. and information customization services. The core of this concept is that users can choose and customize the information content they want to receive based on their interests, preferences and values. Nearly 30 years have passed, "My Daily" is no longer a dream. Whether it is domestic Toutiao, Douyin, Kuaishou, or foreign social media such as TikTok and Facebook, they have integrated a "personalized recommendation" mechanism, and users can easily Effortlessly get recommendations that match your taste. After the birth of generative AI, people are trying to further optimize the process of obtaining information and directly generate their own "daily report". The AI ​​search engine Perplexity was originally created for this purpose. Founder Aravind Srinivas has interned at OpenAI, DeepMind and Google. His first job after graduation was to study language models and diffusion models at OpenAI. In his daily work and life, Aravind found that Google searches sometimes did not directly give him the results he wanted, and due to the influence of Google's advertising business, the links given in the search results were always mixed with advertising links. In 2022, with the development of generative AI technology, Aravind keenly realized that it might be possible to use AI to recreate a search engine that can directly generate answers. So, in August 2022, Aravind Srinivas and two other founders, Denis Yarats and Johnny Ho, established Perplexity AI. After users ask questions on Perplexity, the platform will not provide individual web links, but more readable answers compiled and summarized by AI, and the original text that AI has learned from will be attached. Link. With the enthusiastic pursuit of capital, Perplexity AI has already conducted 5 rounds of financing in just two years since its establishment, with a total financing amount of nearly 850 million U.S. dollars, and its current valuation is approximately 9 billion U.S. dollars (approximately RMB 64 billion). Following this, Perplexity gradually began to explore new functions that can be linked to business. The shopping function of Buy with Pro is a new starting point. Function onlineOn the same day, founder Aravind tweeted to introduce the feature: "We are very happy that Perplexity can start with daily shopping and gradually transition from providing answers to questions to providing AI-native e-commerce transactions."

In fact, the combination of AI + e-commerce/retail has attracted more and more companies and platforms to enter the market. According to a recent survey by international data company IDC, 40% of retailers around the world are currently experimenting with the application of generative artificial intelligence in the sales field, and 21% of retailers are already investing in or using generative artificial intelligence. In June this year, e-commerce search engine company Daydream received a $50 million seed round of financing. The company was established in 2023 and focuses on using artificial intelligence technology to provide a personalized online shopping experience, hoping to solve the problems of inefficiency and information fragmentation encountered by consumers on current e-commerce platforms. Daydream's operating logic also relies on AI technology to analyze users' shopping needs for products and try to understand users' shopping habits, thereby helping users better discover the products they want to buy. For example, when you type "find me a solid-color wool jacket that's perfect for winter," Daydream will automatically help you find related products in the product library. In a sense, after "My Daily" became a reality, "My Mall" to meet personalized shopping needs is also being born.

Will AI be the direction of e-commerce evolution?

Consumers are also calling for AI to be integrated into online shopping. A 2024 survey of 20,000 consumers in 26 countries by information technology services company IBM found that despite carrying smartphones in their pockets, consumers often spend hours searching for suitable products while comparing prices. while assessing quality. More than half of respondents said they would like to use robots or virtual assistants (55%), augmented or virtual reality technology (55%) and artificial intelligence applications (59%) to assist them with their shopping. From offline shopping to online e-commerce to today’s AI e-commerce, people’s demand for shopping has always been—faster speed, less labor-saving processes, and a more comfortable experience. Among them, the first two have greatly improved in the process of switching from offline shopping to online e-commerce; but from an experience point of view, the online shopping experience compared with the consumption experience in offline physical stores, Sometimes it may fall instead of rising. In the past, when purchasing goods in stores, there would be attentive service from shopping guides. You only need to state your preferences and purchasing needs, and the shopping guides will help you find the corresponding products on the shelves. During the process of trying out or trying on the products, the shopping guides will We can also promptly explain to you the price, materials, and product precautions. If you are hesitant between two products, the shopping guide will usually help you analyze which product is more suitable for you. In China, although online e-commerce has not been able to completely replace the real-time recommendation and explanation services of offline shopping guides, there have been many efforts in this direction. 2013Taobao and Tmall have started to experiment with "personalized recommendations". The waterfall content on the home page of the application is a product push based on consumers' previous shopping preferences. Live e-commerce, which became popular around 2018, also relies on the recommendations and explanations of anchors to help consumers understand the products they want to buy more intuitively, thereby achieving a better shopping experience. Overseas, e-commerce's progress in optimizing shopping experience is not as fast as in China. Although overseas consumers have also experienced the transition from offline to online, for a long time they have been more accustomed to purchasing directly from the brand's official website after searching on Google. However, as more and more brands and merchants join online sales, the screening cost spent by consumers in the search process is also getting higher and higher. It's like when you go to a huge online supermarket without a shopping guide to guide you, you can only go from shelf to shelf to find what you want to buy, which is time-consuming and labor-intensive. The solution provided by traditional search engines is to charge brands and merchants. Whoever pays more can be placed in the most conspicuous position. However, this undoubtedly considers the sales needs of brands and merchants, not the purchasing needs of consumers. . In fact, this is why overseas AI search engines such as Perplexity and Daydream are targeting e-commerce business. At present, AI search + e-commerce may be an evolutionary path more suitable for overseas e-commerce. By transforming AI into an online shopping guide, it can identify consumers' preferences and provide personalized services, optimizing consumers' online shopping experience in the form of technology. In addition to higher efficiency, everything is evolving in a more "human" direction. This may be the core of the harmonious coexistence between AI and humans in the future.

Reference articles: 1. AI-Enabled E-Commerce Solutions Market Size, InsightAce Analytic 2. Enter AI Search We launched the world’s first B2B AI search engine Accio, Alibaba International Station 3. Three people are valued at 60 billion, Investment World 4, 2024 Consumer Study: Revolutionize retail with AI everywhere, IBM